IAEA Training in level 1 PSA and PSA applications PSA applications PSA-based evaluation and rating of operational events
Content HISTORICAL BACKGROUND PROCEDURES FOR EVENT EVALUATION USING PSA USE OF PSA FOR EVENT RATING CURRENT ACTIVITIES Slide 2.
OBJECTIVES: WHY DO WE WANT TO USE PSA FOR THIS PURPOSE? To gain insights in NPP weaknesses in relation to the occurred situation To obtain a measure of the event severity To validate and backfit the PSA models Slide 3.
GLOSSARY OF TERMS CDF Core Damage Frequency CCDP Conditional Core Damage Probability IE Initiating Event HRA Human Reliability Analysis ASP US-NRC project on Accident Sequence Precursor analysis LER Licensee Event Report IRS IAEA Incident Reporting System INES International Nuclear Event Scale Slide 4.
HISTORICAL BACKGROUND US-NRC ASP (ACCIDENT SEQUENCE PRECURSOR) Analysis of precursors using probabilistic approaches Started in 1979 at the Oak Ridge Laboratory Applicable to USA commercial plants Use of generic & simplified BWR and PWR PSA models Used to evaluate safety significance of the events reported through the LER programme: Calculation of CCDP due to each event CCDP was used as a means of ranking events Slide 5.
HISTORICAL BACKGROUND US-NRC ASP (ACCIDENT SEQUENCE PRECURSOR) (Cont.) Important benefits: ASP showed that PSA-based methods are a good approach for evaluation and ranking of events ASP showed that PSA-based evaluation of events is a good approach for making PSA more complete and realistic (Search for events not or poorly predicted in the PSA) Limitations: Use of generic models Use of simplified models Slide 6.
HISTORICAL BACKGROUND OTHER NATIONAL PROJECTS: GERMANY, SWEDEN, FINLAND, FRANCE Several national projects have developed since the mid eighties i. e. German Precursor Study (1985): To validate German Risk Study using the operational experience of the Biblis NPP: CDFoperexp < CDFrisk study. Slide 7.
HISTORICAL BACKGROUND OTHER NATIONAL PROJECTS: GERMANY, SWEDEN, FINLAND, FRANCE (Cont.) References: "Precursor Studies", H. Hoertner and P. Kafka, published in IAEA-TECDOC-387 "Combining Risk Analysis and Operating Experience", 1986. "Demonstration Case Studies on Living PSA", SKI Technical Report 93:33, NKS/SIK-1 (92)27, August 1993. "Safety Evaluation by Living Probabilistic Safety Assessment. Procedures and Applications for Planning of Operational Activities and Analysis of Operating Experience", SKI Report 94:2, January 1994. "TVO II - Risk Follow-up study of precursors and component failures leading to LCO's, Jan 1985 - May 1994", Finnish Centre for Radiation and Nuclear Safety (STUK) Nuclear Safety Department, Draft Report, October 1994. "Incidents Analysis", Ph. Francois (EdF), Topic 4, International Seminar, November 21-23, 1994. Slide 8.
PROCEDURES FOR EVENT EVALUATION USING PSA BACKGROUND IAEA-TECDOC-611: "Use of plant specific PSA to evaluate incidents at nuclear power plants" (1991) Initiated in the framework of the IAEA-IRS project: optimising experience feedback requires selection of events with higher safety significance Purpose: To characterise the relative importance of incidents Starting point: Use of plant-specific PSA in order to: allow better understanding of NPP vulnerabilities check the PSA model for appropriateness and completeness Slide 9.
PROCEDURES FOR EVENT EVALUATION USING PSA METHODOLOGY Selection of incidents for analysis Understanding of the incident and its safety implications Establishment of the relationship between the incident and the PSA model Modification of the models to reflect the incident Calculation of new PSA results Analysis of the results Slide 10.
PROCEDURES FOR EVENT EVALUATION USING PSA METHODOLOGY (Cont.) Selection of incidents for analysis Initial qualitative screening to select those of most value to the analysis: Any incident which degrades safety functions is a candidate for selection: Incidents that involve safety function failure or degradation Any incident which results in unexpected or significant challenges to the safety functions is a candidate for selection: Initiating events and precursors to initiating events Events occurring at a frequency greater than the estimated frequency for the PSA Multiple failures or degradations Events that are not well modelled in the PSA Slide 11.
PROCEDURES FOR EVENT EVALUATION USING PSA METHODOLOGY (Cont.) Understanding of the incident and its safety implications It requires knowledge of the NPP design and operation in order to determine if the incident impacted or had the potential to impact a safety function It requires knowledge of the contents of the PSA to determine if the potential impacts are within the scope or resolution of the PSA models Slide 12.
PROCEDURES FOR EVENT EVALUATION USING PSA METHODOLOGY (Cont.) Establishment of the relationship between the incident and the PSA model To determine: Accident sequences involved Fault trees and basic events that model the components and operator actions concerned Which recovery actions could be claimed and which are impossible Slide 13.
PROCEDURES FOR EVENT EVALUATION USING PSA METHODOLOGY (Cont.) Modification of the models to reflect the incident i.e.: Restoring accident sequences that were originally truncated out in the final results Changing basic event probabilities Evaluating new human error rates Slide 14.
PROCEDURES FOR EVENT EVALUATION USING PSA METHODOLOGY (Cont.) Calculation of new PSA results Calculation of the new results conditional on the existence of the incident: Events that could have lead to an initiating event but no IE occurred Initiating events Events that jeopardise NPP response to an initiating event but no IE occurred Slide 15.
PROCEDURES FOR EVENT EVALUATION USING PSA METHODOLOGY (Cont.) Analysis of the results Comparison of new and existing results Determination of the new dominant contributors Determination of the new importance of the remaining systems, components and operator actions to prevent core damage Slide 16.
PROCEDURES FOR EVENT EVALUATION USING PSA DETAILED PROCEDURE Review of the incident Identification of all the event tree sequences affected by the incident Review of the identified sequences and cut-sets to determine if the affected systems and basic events were not truncated in the original results Determination of the best estimate failure probabilities for all basic events impacted by the incident Calculation of the new accident sequence conditional probability Determination of the recovery actions to be applied to the cut-sets Calculation of importance measures for the new results Performance of the required sensitivity analyses Documentation of the analysis Slide 17.
PROCEDURES FOR EVENT EVALUATION USING PSA DETAILED PROCEDURE (Cont.) Review of the incident Identify the chronology of events Identify all equipment failures, degradations and unavailabilities Note all operator actions taken, specially those not covered by procedures and training Review related problems and conditions that occurred a period before and after the event Slide 18.
PROCEDURES FOR EVENT EVALUATION USING PSA DETAILED PROCEDURE (Cont.) Identification of all the event tree sequences affected by the incident The full event tree models should be used The analyst must know which event tree headings include the equipment and operator actions involved in the event The sequences of concern are those with a failure branch for at least one of the headings of concern Slide 19.
PROCEDURES FOR EVENT EVALUATION USING PSA DETAILED PROCEDURE (Cont.) Review of the identified sequences and cut-sets to determine if the affected systems and basic events were not truncated in the original results If the sequences or cut-sets of concern were not retained, they have to be generated. The cut-off criteria should be properly selected Slide 20.
PROCEDURES FOR EVENT EVALUATION USING PSA DETAILED PROCEDURE (Cont.) Determination of the best estimate failure probabilities for all basic events impacted by the incident Failed components can be modelled as house events "1" to correctly generate the Boolean logic Unavailabilities: their duration must be taken into account: Multiplying the IE frequency by the unavailability time (fraction of the year) Introducing the real unavailability event (instead if a failed house event) Equipment or operator degradations: Detailed analyses may be required to calculate the new failure probabilities Slide 21.
PROCEDURES FOR EVENT EVALUATION USING PSA DETAILED PROCEDURE (Cont.) Calculation of the new accident sequence conditional probability Events that could have lead to an initiating event but no IE occurred: Development of new event trees may be required Initiating events: calculation of the Conditional Core Damage Probability with the NPP boundary conditions at the time of the incident (based on the NPP status) Events that jeopardise NPP response to an initiating event but no IE occurred: calculation of all Conditional Core Damage Probabilities for all IEs for which plant response could be degraded by the incident WARNING: Impossible combinations Slide 22.
PROCEDURES FOR EVENT EVALUATION USING PSA DETAILED PROCEDURE (Cont.) Determination of the recovery actions to be applied to the cut-sets Based on the events of the incident Available personnel should be considered The operating procedures should be taken into account Dedicated HRA may be require Recovery actions included in the original PSA should be reconsidered Slide 23.
PROCEDURES FOR EVENT EVALUATION USING PSA DETAILED PROCEDURE (Cont.) Calculation of importance measures for the new results Fussell-Vesely Risk Reduction Worth Risk Achievement Worth Slide 24.
PROCEDURES FOR EVENT EVALUATION USING PSA DETAILED PROCEDURE (Cont.) Documentation of the analysis The documentation should be clear concise and traceable All the steps of the process should be fully documented Slide 25.
USE OF PSA FOR EVENT RATING OBJECTIVE To develop a consistent measure of the relative safety significance of each event Slide 26.
USE OF PSA FOR EVENT RATING WARNINGS The relative significance of an event might be different depending on the "risk index" used, i.e. Core Damage Frequency vs. Frequency of Large Release A PSA-based approach to rate events will not be useful for all the events that happen at the plant (i.e. radiation exposure, waste production). Care must be taken when comparing the safety significance of all NPP events that have safety implications Slide 27.
USE OF PSA FOR EVENT RATING PROPOSALS First proposal: Measures of importance Events can be rated based on the importance or weight of the relevant cut-sets. The following approach could be followed: Modification of the event/s frequency/ies according to the observations Calculation of the new CDF due to the impacted sequences (CDF c ) Normalisation of the obtained CDF c with respect to the baseline CDF Slide 28.
USE OF PSA FOR EVENT RATING PROPOSALS (Cont.) Second proposal: PSA direct results Operational events could be rated according to the Probability of Core Damage conditional to the occurrence of the event Slide 29.
USE OF PSA FOR EVENT RATING CONCLUSIONS Whatever risk-based method, existing or newly developed, is used for event rating: The measures developed should be comparable for all events analysed The measures should be calculated on a consistent basis The rating scale should be understood, by the analysts that perform the analyses and by the organisations who might receive rating reports and possibly base decisions on these Slide 30.
EXAMPLES OF CURRENT ACTIVITIES RISK-BASED APPROACH TO ANALYSING OPERATING EVENTS Ref: Marchese, A.R., Neogy, P. Risk-based approach to analysing operating events. Proceedings of Probabilistic Safety Assessment and Management 96 (ESREL 96 PSAM-III), Crete, Greece, June 24-28, 1996. This work was supported by the US Department of Energy Slide 31.
EXAMPLES OF CURRENT ACTIVITIES RISK-BASED APPROACH TO ANALYSING OPERATING EVENTS (Cont.) The proposed risk-based approach for evaluation of operational events consists of the following steps: Qualitative estimations of conditional probability and consequence -> qualitative estimation of risk Comparison to risk criteria Semi-quantitative estimations of conditional probability and consequence -> semi-quantitative estimation of risk (using simplified models) Comparison to reference value Refinement of the risk estimate Cost-benefit analysis of possible system and procedure modifications Slide 32.
EXAMPLES OF CURRENT ACTIVITIES PROBABILISTIC INCIDENT ANALYSIS (EDF, FRANCE) Ref: Dubreuil-Chambardel A., Francois P., Pesme H., Maliverney B., An Operating PSA Application at EDF: the Probabilistic Incident Analysis. Proceedings of Probabilistic Safety Assessment and Management 96 (ESREL 96 PSAM-III), Crete, Greece, June 24-28, 1996. Slide 33.
EXAMPLES OF CURRENT ACTIVITIES PROBABILISTIC INCIDENT ANALYSIS (EDF, FRANCE) (Cont.) Objective Identification and analysis of all the incidents that might degenerate and result in core damage or radioactive release Slide 34.
EXAMPLES OF CURRENT ACTIVITIES PROBABILISTIC INCIDENT ANALYSIS (EDF, FRANCE) (Cont.) Systematic selection of events Qualitative criteria based on safety significance Slide 35.
EXAMPLES OF CURRENT ACTIVITIES PROBABILISTIC INCIDENT ANALYSIS (EDF, FRANCE) (Cont.) Analysis Calculation of conditional probability of core meltdown Slide 36.
EXAMPLES OF CURRENT ACTIVITIES PROBABILISTIC INCIDENT ANALYSIS (EDF, FRANCE) (Cont.) Definition of accident precursor Is any incident whose conditional probability of core melt is greater that 1E-6 when all the parameters are adjusted Slide 37.
EXAMPLES OF CURRENT ACTIVITIES PROBABILISTIC INCIDENT ANALYSIS (EDF, FRANCE) (Cont.) Key points Promotion of a risk-analysis culture among operators Value of accident feedback Understanding of how serious it is and why it is serious Slide 38.
OTHER REFERENCES "The USNRC Accident Sequence Precursor Program: Present Methods and Findings", J.W. Minarick, Reliability Engineering and System Safety 27 (23-51), 1990. "Precursors to Potential Severe Core Damage Accidents: 1969-1979, A Status Report." NUREG/CR-2497, vols 1 and 2, J.W. Minarick and C.A. Kukielka, Oak Ridge National Laboratory, 1982. "Accident Sequence Precursor Program Methods", Presentation by J.W. Minarick, NUREG CP-0124, Proceedings of the Workshop on the Use of PRA Methodology for the Analysis of Reactor Events and Operational Data - held at Loew's Hotel, Annapolis, Maryland, January 29-30, 1992. Slide 39.
OTHER REFERENCES (Cont.)."Risk Assessment in Event Evaluations Training Course", Idaho National Engineering Laboratory, USNRC, Washington DC, June 1995.."Precursors to Potential Severe Core Damage Accidents: 1992 A Status Report", NUREG/CR-4674 (ORNL/NOAC-232), vol. 17 and 18, December 1993..Boneham, P. The benefits of using PSA to enhance the feedback of operational experience at Nuclear Power Plants, COPSA 97, Edinburgh, U.K. October 97.Hadnagy, L. Use of PSA at Paks NPP - PSA based event evaluation To be presented in the Technical Committee Meeting on PSA Applications to improve NPP safety, Madrid, Spain, 23-27 February, 1998.Hernández-Arteaga Use of Laguna Verde PSA to Analyze the ECCS Strainer Blockage Issue To be presented in the Technical Committee Meeting on PSA Applications to improve NPP safety, Madrid, Spain, 23-27 February, 1998 Slide 40.